Most AI tools make you explain the context before they can help. Goldfish already has it. It privately remembers what you’ve been working on across your Mac, then helps you write better from any app. Press Option in a text field to draft replies, summarize threads, rewrite sentences, or recall important details from your recent work without copying, pasting, or re-explaining the whole backstory.
I have a serious question - you take over someones computer, read their screen, pass it to ai or other service (who knows) and there is ZERO information on your company who is responsible for that software, is it compliant with any software standard. Dont you think its getting a little bit too crazy? I am aware that young people dont care about the privace and what can go wrong? Right?
Goldfish
@igor_gnot1 Hi Igor! We have our privacy policy here: https://www.goldfish.sh/privacy, which is shown during the onboarding, as well as our TOS here: https://www.goldfish.sh/terms
In short, your memory stays on your Mac in an encrypted local database. Goldfish reads text through macOS Accessibility, not screenshots or screen recordings. Text only leaves your device when you actively invoke Goldfish, and then only the relevant context for that request is sent to our private Azure OpenAI deployment with zero data retention.
Thank you for the feedback this might not be clear enough, we should probably also disclose how to get in touch with us to further increase the trust in Goldfish 🐠
@igor_gnot1 i understand your worry and we will improve our communication as we grow. for now you'll have to trust us and the very generous documentation that exists online. Any specific questions and I'm happy to answer! :)
Does it work equally well across different apps like Slack, Gmail, and Notion? The feature sounds great, but I'm curious about how seamless the integration really is.
Goldfish
@neri_zavadil It does! There are some small differences between apps, because every text field behaves a bit differently, but the goal is exactly that it should work across Slack, Gmail, Notion, and basically anywhere you can type.
If you try it and find a surface where it feels rough, I’d genuinely love to hear which one.
@neri_zavadil yes, that’s the idea. The tricky part is making each surface feel native, because Slack, Gmail, Notion etc all expose text fields a bit differently.
what are your best use cases so far? 🐠
@joel_edholm Congrats on your Goldfish launch! Curious about privacy and data storage—specifically, whether Goldfish processes our context locally on the device or if it sends any data to the cloud to mimic our style, our writing, etc.?
Goldfish
@joel_edholm @wai_kit_cheah Thanks Wai Kit. Goldfish uses local memory on your Mac, not cloud sync. The context it captures is stored in a local database, and there is no backend where we can browse your data.
When you ask it to write or answer something, the relevant context is sent to the AI provider for that request, but with zero data retention. You can also pause capture whenever you want, exclude specific apps or domains, and delete stored history.
@wai_kit_cheah fair question. Today Goldfish uses cloud models for the actual generation, but the context is scoped to what you’re doing and we’re very intentional about what gets sent, stored, and remembered. Local-first/on-device is the direction we want to push as the models get good enough.
what are your best use cases so far? 🐠
@joel_edholm Local-first/on-device is a good direction.
I would be interested on how Goldfish could handle context switching. Let me elaborate. While it's good to have Goldfish write, think, work like us, but if i am dealing with multiple internal stakeholders, or a sales exec managing umpteen accounts, or consultant supporting multiple clients, we would have to remember each of their specific pain points / deliverables / context. And a lot of these details spread across multiple apps, data sources, messages, emails, etc. If Goldfish is local, could it also act like a 'context router', i.e. not just draft a reply, but pick up details/info from Slack, from project plans on OneDrive, or chat conversations I had with the client?
@joel_edholm hello, great app tbh. but i have 1 ques, is there a way to disable the orange line in the top of the screen and if i want to talk to goldfish i press on it directly? thanks in advance
Goldfish
@joel_edholm @itzw0lf You can turn on auto-hide in settings. Then just press Goldfish when you want it to appear on your desktop.
@itzw0lf hope you found it! What have been your main use cases?
Great tool, love it. One thing I found annoying is that the tool sits at the top of the screen, so when I try to switch tabs it blocks me from switching tabs. Also, auto-hide top bar doesn’t seem to be working on Windows.
Another concern I have is what the limitations will be after the beta.
Goldfish
@bojan_inkyy Thanks Bojan, super useful feedback. We haven't unfortunately implemented auto-hide on Windows yet, still lots of improvements on that side of the pond.
On beta limits, we haven’t finalized the exact packaging yet, but the Product Hunt early access includes 3 months free access. We’ll be clear about limits before anything changes! :-D
@bojan_inkyy Sorry about the Windows issues! We'll get on it!
Since Goldfish is sitting across everything on the Mac messages, docs, tabs, meetings, how are you handling the privacy side of that? Is everything processed locally or does any of it leave the device when it's helping you draft something?
Goldfish
@ana_popescu2 Privacy is core to this. The memory stays on your Mac in a local database, with no cloud sync and no backend where we can browse your data.
When Goldfish helps draft something, it uses the relevant local context, and AI calls have zero data retention. You can also pause capture whenever you want, exclude specific apps or domains, and delete stored history.
AI has memory like a goldfish is such a good way to put it. The real tax of every AI tool right now isn't the thinking, it's the re-explaining. How does Goldfish decide what's actually relevant context vs. just noise from everything open on the Mac? Feels like that filtering is the hard part.
Goldfish
@andrew_paul11 We think about it less as “throw everything into memory” and more as ranking a few layers separately.
The current surface wins first: the field you’re typing in, the thread around it, the app/window you’re in. Then recent activity is used for handoffs, like if you just switched from a doc or Slack thread into an email reply. Longer-term memory is mostly for patterns around people, projects, tone, and recurring work.
If Goldfish is unsure, it should stay general or ask, not confidently pull in some random tab from earlier. That filtering is definitely the hard part. More memory is easy, choosing the right memory is the product.